What is your scientific background?

I graduated from the Israel Institute of Technology with a bachelor’s degree in industrial engineering and a master’s degree in operations research, summa cum laude. My graduate thesis was a study of expected call center waiting times of impatient customers. This work was conducted under the supervision of Professor A. Mandelbaum. My first position out of school was as an algorithm developer at Earnix. I remember my first project, very well. It was the implementation of generalized additive model (GAM) regressions. It was during those first months there I learned a lot about statistics, programming, optimization, and the application of “academic” ideas in industry.

My next roles were outside of Israel when my family and I lived abroad in Russia. I had several positions in analytics ranging from data analyst at Nielsen, the global leader in audience measurement, to head of analytics in the customer experience department for VimpelCom Russia, one of the largest Russian telecom operators.

When we returned to Israel eight years ago my first thought was to see if Earnix had a role that fit. I was pleased to find an open position that not only met the goals of my next career move, but also afforded me the opportunity to return to the environment where I had learned so much. It is very exciting that in my current position, I can see very different aspects of the banking and insurance industries. 

What are you working on at the moment? How do you think this work will make a difference?

My day-to-day research focuses on investigating how AI can be leveraged to analyse data and exploring Machine Learning (ML) techniques to assist Earnix customers. The rise of powerful AI technologies is one of the most interesting research topics applied directly to the world of insurance and banking today. I also work a lot on different optimization challenges to determine use cases for insurance companies and banks. 

One of the most interesting projects I participated in at Earnix is its new modelling application, Automatic GLM (AGLM). It combines the power of both “old school” statistical modelling with the new ML approaches and automates the hardest step in creating a strong GLM model. By applying different ML techniques behind the scenes, it manages to automatically complete feature selection and feature engineering. This ensures that the initial GLM model is a very good starting point for any developer. It saves days if not weeks of manual work. I love the way advanced modelling techniques do all the heavy lifting and accelerate time to market for companies working in highly competitive industries, with very fine profit margins and a lot of potential unpredictability when assessing risk. This is an area Earnix excels in, and I enjoy working at the epicentre.

Besides your scientific interests, what are your personal interests?

I have a busy family life raising two girls. I especially enjoy reading, sewing, and spending time with my daughters on different STEM projects. I hope I can show them how exciting science is and that they take on the love of it from me; and the drive for excellence from their dad to grow up to be happy and successful.

In your opinion, which changes, if any, are needed in the scientific system to be more attractive to women in science and possible future scientists?

There are many talented women making significant contributions to the fields of AI, ML and data science. Their expertise, research and innovative solutions demonstrate the crucial role women play in shaping the future of technology. AI is currently one of the most talked-about fields in technology, but unfortunately, there is still a significant lack of female representation in the field.

A recent article in MIT Sloan Management Review points out that it is mathematically impossible to solve for fairness. Naturally, the more diverse a workforce over time, the more biases can be reduced. This is why it is so important to encourage more women to enter the AI field. 

For anyone still daunted by the thought, I would say that most of the resources to become an ML scientist are available online and studying does not require any special equipment except your computer.  Especially today, using AI, you too can learn as you go along and adapt accordingly, learning from what has happened in the past. I am not saying that it is easy. Of course, it requires time and commitment, but it’s an exciting ever-changing field to work in and offers a potentially very lucrative and rewarding career path.